Comparative studies of model perturbations for ICON-D2-EPS.

Convection-permitting ensemble forecasting is well established as a tool to support the prediction of severe weather and high-impact weather in several meteorological centres. At the Deutscher Wetterdienst, ICON-D2-EPS is the limited-area high-resolution component of the ICON modeling system, running as an ensemble of 20 members at 2 km horizontal resolution over Germany and surrounding areas. Several products to support operational forecasting and warning issuing are regularly issued. The perturbed initial conditions are provided by the km-scale ensemble data assimilation system KENDA, run at the same resolution, assimilating a wide range of observations, included radar-derived radar volumes. Boundary conditions are provided by ICON-EPS, the global ensemble with a refinement at 20 km on Europe. Simple model perturbations are applied by Parameter Perturbation, where a set of parameters of the physics parametrisation schemes are perturbed with respect to the default value. In this work, the effect of the model perturbation is studied, by assessing how it contributes to generate diversity between the members for case studies, as compared to a standard operational verification carried out on periods, and how it influences the ensemble spread, in addition to the initial and boundary condition perturbations. The extent to which the ensemble spread is able to represent forecast uncertainty is discussed for near-surface weather parameters. The effect of increasing the amplitude of the parameter perturbation is also addressed. Finally, the impact of the parameter perturbation is discussed in comparison with physically-based perturbation methods which have been implmented and tested in ICON thanks to collaborative works.